U.S. patent application number 14/680705 was filed with the patent office on 2015-07-30 for method and apparatus for optimizing the evaluation of semantic web queries.
The applicant listed for this patent is International Business Machines Corporation. Invention is credited to Mihaela Ancuta Bornea, Julian Timothy Dolby, Anastasios Kementsietsidis, Kavitha Srinivas.
Application Number | 20150213089 14/680705 |
Document ID | / |
Family ID | 51655226 |
Filed Date | 2015-07-30 |
United States Patent
Application |
20150213089 |
Kind Code |
A1 |
Bornea; Mihaela Ancuta ; et
al. |
July 30, 2015 |
Method and Apparatus for Optimizing the Evaluation of Semantic Web
Queries
Abstract
A semantic query over an RDF database is received with RDF
database statistics and access methods for evaluating triple
patterns in the query. The semantic query is expressed as a parse
tree containing triple patterns and logical relationships among the
triple patterns. The parse tree and access methods create a data
flow graph containing a plurality of triple pattern and access
method pair nodes connected by a plurality of edges, and an optimal
flow tree through the data flow graph is determined such that costs
are minimized and all triple patterns in the semantic query are
contained in the optimal flow tree. A structure independent
execution tree defining a sequence of evaluation through the
optimal flow tree is created and is transformed into a database
structure dependent query plan. This is used to create an SQL query
that is used to evaluate the semantic query over the RDF
database.
Inventors: |
Bornea; Mihaela Ancuta;
(White Plains, NY) ; Dolby; Julian Timothy;
(Bronx, NY) ; Kementsietsidis; Anastasios; (New
York, NY) ; Srinivas; Kavitha; (Rye, NY) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
International Business Machines Corporation |
Armonk |
NY |
US |
|
|
Family ID: |
51655226 |
Appl. No.: |
14/680705 |
Filed: |
April 7, 2015 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13856428 |
Apr 3, 2013 |
9031933 |
|
|
14680705 |
|
|
|
|
Current U.S.
Class: |
707/718 |
Current CPC
Class: |
G06F 16/9024 20190101;
G06F 16/3331 20190101; G06F 16/24544 20190101; G06F 16/2453
20190101; G06F 16/24522 20190101; G06F 16/24542 20190101 |
International
Class: |
G06F 17/30 20060101
G06F017/30 |
Claims
1. A method for optimizing semantic web queries, the method
comprising: receiving inputs, the inputs comprising: a semantic
query over a resource description framework database, the semantic
query comprising a plurality of hierarchically nested graph
patterns, each graph pattern comprising at least one triple pattern
comprising a subject, an object and predicate connecting the
subject to the object; statistics describing data in the resource
description framework database; and a plurality of access methods
comprising alternative methods for evaluating each triple pattern;
expressing the semantic query as a query parse tree comprising a
plurality of nodes, the nodes comprising triple patterns from the
query and logical relationships among the triple patterns; creating
a data flow graph, the data flow graph comprising a plurality of
triple pattern and access method pair nodes connected by a
plurality of edges representing dependencies among execution of the
triple patterns; determining an optimal flow tree through the data
flow graph, the optimal flow comprising a set of triple pattern and
access method pair nodes and edges such that all triple patterns in
the semantic query are contained in the optimal flow tree; creating
an execution tree independent of a structure of the resource
description framework database, the execution tree comprising a
sequence of evaluation of triple patterns in the optimal flow tree;
transforming the execution tree into a query plan that exploits the
structure of the resource description framework database; using the
query plan to create a structured query language query; using the
structured query language query to evaluate the semantic query over
the resource description framework database.
2. The method of claim 1, wherein the semantic query comprises a
SPARQL query and the query plan comprises a SPARQL query plan.
3. The method of claim 1, wherein the statistics comprise a total
number of triples in the resource description framework database,
average number of triples per subject, average number of triples
per object, an identification of at least one of uniform resource
identifiers and literals that appear in a greatest number of
triples or combinations thereof.
4. The method of claim 1, wherein the plurality of access methods
comprise access a triple patters by subject, access a triple
pattern by object, conduct a full scan or combinations thereof.
5. The method of claim 1, wherein the logical relationships among
triple patterns in the query parse tree comprise at least one of
AND, OR, UNION and OPTIONAL.
6. The method of claim 1, wherein each edge in the plurality of
edges in the data flow graph denotes a first triple pattern,
producing a variable required in either an object or a subject of a
second triple pattern.
7. The method of claim 1, wherein creating the data flow graph
further comprises: defining a root node in the data flow graph;
selecting an access method from the plurality of access methods;
identifying a triple pattern having a constant in at least one of
the subject and the object that can be accessed by the selected
access method and lacking any required variable, a required
variable comprising a variable in the triple pattern that is
required to be bound when evaluated by the selected access method
by a prior triple pattern evaluation; and connecting a first triple
pattern and access method pair to the root node comprising the
selected access method and the identified triple pattern.
8. The method of claim 7, wherein creating the data flow graph
further comprises: connecting subsequent triple pattern and access
method pairs to the first triple pattern and access method pair by
identifying a required variable in a subsequent triple pattern that
is required to be bound by evaluation of the identified triple
pattern when the subsequent triple pattern is evaluated by a
subsequent selected access method.
9. The method of claim 8, wherein creating the data flow graph
further comprises: calculating a cost of evaluating for a
evaluating a given triple pattern by a given access method with
respect to the statistics of the resource description framework
database; and assigning the calculated cost to edges in the data
flow graph associated with triple pattern and access pair nodes
comprising the given triple pattern and the given access
method.
10. The method of claim 1, wherein: each edge in the data flow
graph comprises a cost associated with evaluating a given triple
pattern by a given access method with respect to the statistics of
the resource description framework database for a triple pattern
and access pair node associated with that edge; and the step of
determining the optimal flow tree further comprises determining and
optimal flow tree such that the edges in the optimal flow tree
comprise a minimum cumulative cost.
11. The method of claim 1, wherein creating the execution tree
further comprises: evaluating contents of each node in the parse
tree sequentially starting with a root node in the parse tree for
inclusion in the execution tree; and determining whether the
contents of a given node are to be added to the execution tree
immediately or held for inclusion in the execution tree after
contents of other nodes are added to the execution tree.
12. The method of claim 11, wherein determining whether the
contents of a given node are to be added immediately further
comprises adding contents of each node at a latest possible point
in the execution tree upon determination that variables or
operators in contents are required for subsequent stages in the
execution tree.
13. The method of claim 1, wherein transforming the execution tree
into a query plan further comprises merging triple patterns having
at least one of a common subject and common object into a single
node in the query plan.
14. The method of claim 13, wherein: the query plan comprises a
plurality of nodes; and using the query plan to create the
structured query language query comprises creating a structured
query language query for each node on the query plan.
15. The method of claim 1, the method further comprises
communicating results of the structured query language query
evaluation of the semantic query.
16. A computer-readable storage medium containing a
computer-readable code that when read by a computer causes the
computer to perform a method for optimizing semantic web queries,
the method comprising: receiving inputs, the inputs comprising: a
semantic query over a resource description framework database, the
semantic query comprising a plurality of hierarchically nested
graph patterns, each graph pattern comprising at least one triple
pattern comprising a subject, an object and predicate connecting
the subject to the object; statistics describing data in the
resource description framework database; and a plurality of access
methods comprising alternative methods for evaluating each triple
pattern; expressing the semantic query as a query parse tree
comprising a plurality of nodes, the nodes comprising triple
patterns from the query and logical relationships among the triple
patterns; creating a data flow graph, the data flow graph
comprising a plurality of triple pattern and access method pair
nodes connected by a plurality of edges representing dependencies
among execution of the triple patterns; determining an optimal flow
tree through the data flow graph, the optimal flow comprising a set
of triple pattern and access method pair nodes and edges such that
all triple patterns in the semantic query are contained in the
optimal flow tree; creating an execution tree independent of a
structure of the resource description framework database, the
execution tree comprising a sequence of evaluation of triple
patterns in the optimal flow tree; transforming the execution tree
into a query plan that exploits the structure of the resource
description framework database; using the query plan to create a
structured query language query; using the structured query
language query to evaluate the semantic query over the resource
description framework database.
17. The computer-readable storage medium of claim 16, wherein
creating the data flow graph further comprises: defining a root
node in the data flow graph; selecting an access method from the
plurality of access methods; identifying a triple pattern having a
constant in at least one of the subject and the object that can be
accessed by the selected access method and lacking any required
variable, a required variable comprising a variable in the triple
pattern that is required to be bound when evaluated by the selected
access method by a prior triple pattern evaluation; connecting a
first triple pattern and access method pair to the root node
comprising the selected access method and the identified triple
pattern; and connecting subsequent triple pattern and access method
pairs to the first triple pattern and access method pair by
identifying a required variable in a subsequent triple pattern that
is required to be bound by evaluation of the identified triple
pattern when the subsequent triple pattern is evaluated by a
subsequent selected access method.
18. The computer-readable storage medium of claim 16, wherein: each
edge in the data flow graph comprises a cost associated with
evaluating a given triple pattern by a given access method with
respect to the statistics of the resource description framework
database for a triple pattern and access pair node associated with
that edge; and the step of determining the optimal flow tree
further comprises determining and optimal flow-tree such that the
edges in the optimal flow tree comprise a minimum, cumulative
cost.
19. The computer-readable storage medium of claim 16, wherein
creating the execution tree further comprises: evaluating contents
of each node in the parse tree sequentially starting with a root
node in thee parse tree for inclusion in the execution tree; and
determining whether the contents of a given node are to be added to
the execution tree immediately or held for inclusion in the
execution tree after contents of other nodes are added to the
execution tree; wherein determining whether the contents of a given
node are to be added immediately further comprises adding contents
of each node at a latest possible point in the execution tree upon
determination that variables or operators in contents are required
for subsequent stages in the execution tree.
20. The computer-readable storage medium of claim 16, wherein:
transforming the execution tree into a query plan further comprises
merging triple patterns having at least one of a common subject and
common object into a single node in the query plan, the query plan
comprising a plurality of nodes; and using the query plan to create
the structured query language query comprises creating an
structured query language query for each node on the query plan.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The present application is a continuation of co-pending U.S.
patent application Ser. No. 13/856,428 filed Apr. 3, 2013. The
entire disclosure of that application is incorporated herein by
reference.
FIELD OF THE INVENTION
[0002] The present invention relates to semantic web queries.
BACKGROUND OF THE INVENTION
[0003] Resource Description Framework (RDF) is the de-facto
standard for graph representation and the primary vehicle for data
exchange over the Internet or World Wide Web. RDF is flexible and
uses simple primitives for data representation, e.g., nodes and
edges. In addition, RDF facilitates the integration of
heterogeneous sources on the Web. The query language of choice for
RDF is SPARQL. SPARQL queries are complex and contain a large
number of triples and several layers of nesting. Optimization of
SPARQL queries involves defining the order and methods with which
to access the triples and building a hierarchical plan tree for
query evaluation based on cost. A number of works have already
studied how to efficiently evaluate semantic web (SPARQL) queries.
Typical existing approaches are performing bottom-up SPARQL query
optimization, i.e., individual triples or conjunctive patterns in
the SPARQL query are independently optimized and then each
optimizer attempts to piece together and order these individual
plans into one global plan. These approaches are similar to typical
relational database optimizers in that they rely on statistics to
assign costs to query plans and are in contrast to less effective
approaches whose SPARQL query optimization heuristics ignore
statistics.
[0004] Simple SPARQL queries resemble Structured Query Language
(SQL) conjunctive queries, and, therefore, one expects that
existing techniques to be sufficient. However, a simple overview of
real and benchmark SPARQL queries shows that SPARQL queries
encountered in practice are far from simple. To a large extent due
to the nature of RDF, these SPARQL queries are often arbitrarily
complex, e.g., with deep nestings, and often, quite big, e.g. one
exemplary SPARQL query involves a union of 100 queries. To make
matters worse, typical operators in SPARQL often correspond to more
exotic operators in the relational world that are less commonly
considered by optimizers. For example, the common OPTIONAL operator
in SPARQL corresponds to left-outer joins. All these observations
lead to the conclusion that there is potential for novel
optimization techniques in this space.
[0005] Although attempts have been made to provide query
optimization both in SPARQL and beyond, important challenges remain
for SPARQL query optimization translation of SPARQL queries to
equivalent SQL queries over a relational database or store. Typical
approaches perform bottom-up SPARQL query optimization, i.e.,
individual triples or conjunctive SPARQL patterns are independently
optimized and then the optimizer orders and merges these individual
plans into one global plan. These approaches are similar to typical
relational optimizers in that they rely on statistics to assign
costs to query plans. While these approaches are adequate for
simple SPARQL queries, they are not as effective for more
complicated, but still common, SPARQL queries. Such queries often
have deep, nested sub-queries whose inter-relationships are lost
when optimizations are limited by the scope of single triple or
individual conjunctive patterns.
SUMMARY OF THE INVENTION
[0006] Exemplary embodiments of systems and methods in accordance
with the present invention are directed to a hybrid two-step
approach to query optimization. As a first step, a specialized
structure, called a data flow is constructed that captures the
inherent inter-relationships due to the sharing of common variables
or constants of different query components. These
inter-relationships often span the boundaries of simple conjuncts
(or disjuncts) and are often across the different levels of nesting
of a query, i.e., they are not visible to existing bottom-up
optimizers. As a second step, the data flow and cost estimates are
used to decide both the order with which to optimize the different
query components and the plans that are going to be considered.
[0007] While the hybrid optimizer searches for optimal plans, this
search is qualified by the fact that SPARQL queries are ultimately
converted to SQL. That is, the plans are created such that when
they are implemented in SQL they are amenable to optimizations by
the relational query engine and can be efficiently evaluated in the
underlying relational store. Therefore, SPARQL acts as a
declarative query language that is optimized, while SQL becomes a
procedural implementation language. This dependence on SQL
essentially transforms the problem from a purely query optimization
problem into a combined query optimization and translation problem.
The translation part is particularly complex since there are many
equivalent SQL queries that implement the same SPARQL query
plan.
[0008] The hybrid optimization and the efficient SPARQL-to-SQL
translation are generalizable and can be applied in any SPARQL
query evaluation system. The hybrid optimizer can be used for
SPARQL query optimization independent of the selected RDF storage,
i.e., with or without a relational back-end. The efficient
translation of SPARQL to SQL can be generalized and used for any
relational storage configuration of RDF. The combined effects of
these two independent contributions drive the performance of the
present invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 is a schematic representation of an embodiment of a
query optimization and translation architecture for use in
accordance with the present invention;
[0010] FIG. 2 is an illustration of an embodiment of a sample input
for query optimization;
[0011] FIG. 3 is an illustration of an embodiment of query parse
tree;
[0012] FIG. 4 is an illustration of an embodiment of a data flow
graph;
[0013] FIG. 5 is an illustration of an embodiment of an execution
tree;
[0014] FIG. 6 is an illustration of an embodiment of a query plan
tree;
[0015] FIG. 7 is an illustration of an embodiment of a SQL code
template;
[0016] FIG. 8 is an illustration of an embodiment of results from
performing a translation without using an optimal method of
translation;
[0017] FIG. 9 is a graph illustrating LUBM benchmark results;
[0018] FIG. 10 is a graph illustrating PRBench sample of
long-running queries; and
[0019] FIG. 11 is a graph illustrating PRBench sample of
medium-running samples.
DETAILED DESCRIPTION
[0020] Systems and methods in accordance with the present invention
are directed to a hybrid SPARQL query optimization technique that
is generic and independent of the choice of representing RDF data
in relational schema or otherwise. Therefore, the query
optimization techniques can be used by any other optimizer in this
space. Optimizations that are representation-agnostic are separated
from those that are representation-specific. This modularity
provides significant advantages and the ability to fine-tune the
storage and query optimization layers independently of each
other.
[0021] Exemplary embodiments in accordance with the present
invention achieve huge performance gains by independently
optimizing SPARQL and the SPARQL to SQL. The hybrid SPARQL query
optimization technique is generic and independent of how the RDF
data are represented, e.g., relational schema, or otherwise. In
fact these techniques can be applied directly to query optimization
for native RDF stores. The query translation techniques are then
tuned to our schema representation. Referring initially to FIG. 1,
an exemplary embodiment of the steps of the optimization and
translation process as well as the key structures constructed at
each step 100 is illustrated. A plurality of inputs 102 are
communicated to a data flow builder 104. These inputs include a
semantic web query 106 such as a SPARQL query over an underlying
data set, preferably an RDF dataset, statistics over the underlying
dataset 108 and access methods for accessing the triple patterns in
the query 110. The data flow builder contributes to query
optimization by using the inputs to create a query parse tree 112.
The data flow builder then uses the inputs to create a data flow
graph 104 and to identify an optimal flow tree 116 through the data
flow graph that minimizes costs while traversing all nodes in the
data flow graph. The plan builder then creates an execution tree
118 from the optimal flow tree and a query plan 120 is created in
the SQL builder. The query plan is translated into an SQL query 122
before it is used to execute the query.
[0022] Regarding the three inputs, the SPARQL query, Q, conforms to
the SPARQL 1.0 standard. Therefore, each query Q is composed of a
set of hierarchically nested graph patterns, , with each graph
pattern P.di-elect cons. being, in its most simple form, a set of
triple patterns. The statistics S over the underlying RDF dataset
are defined using types and precision with regard to specific
implementations. Examples of collected statistics include, but are
not limited to, the total number of triples, average number of
triples per subject, average number of triples per object, and the
top-k URIs or literals in terms of number of triples in which they
appear. The access methods, , provide alternative ways to evaluate
a triple pattern t for some pattern P.di-elect cons.. The methods
are system-specific and dependent on existing indexes. For example,
for a system having only subject and object indexes, i.e., no
predicate indexes, the methods would be access-by-subject (acs), by
access-by-object (aco) or a full scan (sc).
[0023] Referring to FIG. 2, a sample set of inputs 200 is
illustrated. These inputs include a sample query Q 202 that
retrieves the people that founded or are board members of companies
in the software industry. For each such company, the query
retrieves the products that were developed by that company, the
company's revenue and, optionally, the number of employees in that
company. The statistics S 204 contain the top-k values 206 for the
constants in the query, e.g., IBM or industry and a count 208 for
each constant value the indicates the frequency of occurrence of
that constant value in the base triples of the query. The access
methods 210 include a set 212 containing a plurality of methods for
accessing the triple patterns in the query. As illustrated, three
different access methods are included in , an access method that
performs a data scan (sc), an access method that retrieves all the
triples given a subject (acs) and an access method that retrieves
all the triples given an object (aco).
[0024] As was shown in FIG. 1, the optimizer includes two modules,
the Data Flow Builder (DFB) 104 and the Query Plan Builder (QPB)
119, Regarding the DFB, triple patterns in the query typically
share variables, and the evaluation of a first query is often
dependent on the evaluation of a second query. Returning to FIG. 2,
for example, triple pattern t.sub.1 214 shares variable ? x 216
with both triple patterns t.sub.2 218 and t.sub.3 220. In the DFB,
sideways information passing is used to construct an optimal flow
tree that considers cheaper patterns, i.e., in terms of estimated
variable bindings, first before feeding these bindings to more
expensive patterns. While the DFB considers information passing
irrespectively of the query structure, i.e., the nesting of
patterns and pattern operators, the QPB module incorporates this
structure to build an execution tree, that is a storage-independent
query plan. The query translation uses this execution tree to
produce a storage specific query plan.
[0025] The DFB starts by building a parse tree for the input query.
Referring to FIG. 3, an embodiment of a parse tree built for the
query in FIG. 2 is illustrated. The parse tree includes a plurality
of nodes 302 for the triples in the query and a plurality of nodes
304 for the relationships among the triples. Sideways information
passing is used with the parse tree to compute the data flow graph
400 illustrated in FIG. 4. The data flow graph takes into account
both a given triple and an access pattern used to access this
triple and represents the dependences among the executions of each
triple pattern. Each node 402 in the data flow graph is a pair
containing a given triple pattern 404 from the query and a given
access method 406 from the input set of access methods. An edge 409
between any two given nodes denotes one triple producing a shared
variable that another triple requires. Starting at any given node
containing a given triple and access method, edges are determined
from the parse tree. Combinations of nodes and edges define a
plurality of routes or paths through the data flow graph that
contain given sets of nodes and edges. Exemplary embodiments of
systems and methods in accordance with the present invention look
for an optimal path through the parse tree that minimizes cost and
that includes all of triples from the original query. Therefore,
the DFB computes the optimal flow tree through the data flow graph
that includes a subset of nodes 408. The optimal flow tree
determines an optimal way in terms of minimizing costs to traverse
all the triple patterns in the query.
[0026] In one embodiment, cost is determined based on a triple
method cost (TMC). The triple method cost is a function that maps
TMC(t,m,S):.fwdarw.c; c.di-elect cons..sub..ltoreq.0, where t is a
given triple, m is an access method, S are the statistics for RDF.
A cost c is assigned to evaluating t using m with respect to
statistics S. The mapping function varies with the degree to which
S are defined. Therefore, the cost estimation depends on the
statistics S. In the example query of FIG. 2, TMC(t.sub.4,aco,S)=2,
because the exact lookup cost using the object of triple t.sub.4,
Software, is known. For a scan method, TMC(t.sub.4,sc,S)=26, i.e.,
the total number of triples in the dataset. Finally,
TMC(t.sub.4,acs,S)=5, i.e., the average number of triples per
subject, assuming subject is bound by a prior triple access.
[0027] Regarding the building of the data flow graph, the data flow
graph models how using the current set of bindings for variables
can be used to access other triples. In modeling this flow, the
semantics of AND, OR and OPTIONAL patterns are respected. A set of
helper functions are used to define the graph. The symbol .uparw.
refers to parents in the query tree structure. For a triple or a
pattern, this is the immediately enclosing pattern. The symbol *
denotes transitive closure. The first helper function is produced
variable. The data flow graph models how using the current set of
bindings for variables can be used to access other triples. In
modeling this flow, the semantics of AND, OR and OPTIONAL patterns
are respected. First, a set of helper functions are introduced and
are used to define the graph. Produced Variables is a function,
(t,m):.fwdarw.V.sub.prod, that maps a triple and an access method
pair to a set of variables that are bound after the lookup, where t
is a triple, m is an access method, and V.sub.prod is the set of
variables in the triple produced by the lookup. In the example
query, for the pair (t.sub.4,aco), P(t.sub.4,aco):.fwdarw.y,
because the lookup uses Software as an object, and the only
variable that gets bound as a result of the lookup is y.
[0028] Required Variables is a function, (t,m):.fwdarw.V.sub.req,
that maps a triple and an access method pair to a set of variables
that are required t be bound for the lookup, where t is a triple, m
is an access method, and V.sub.req is the set of variables required
for the lookup. In the example query, (t.sub.5,aco):.fwdarw.y. That
is, if the aco access method is used to evaluate t.sub.5, then
variable y is required to be bound by some prior triple lookup.
[0029] Least Common Ancestor, LCA(p,p') is the first common
ancestor of patterns p and p'. More formally, it is defined as
follows: LCA(p,p')=xx.di-elect cons..uparw.*(p)x.di-elect
cons..uparw.*(p').E-backward./y.y.di-elect
cons..uparw.*(p)y.di-elect cons..uparw.*(p')x.di-elect
cons..uparw.*(y). As an example, in FIG. 3, the least common
ancestor of AND.sub.N and OR is AND.sub.T. The Ancestors to LCA
function, .uparw..uparw.(p,p') refers to the set of .uparw.* built
from traversing from p to the LCA(p,p'):
.uparw..uparw.(p,p').ident.{x|x.di-elect cons..uparw.*(p)x.di-elect
cons..uparw.*(LCA(p,p'))}
For instance, for the query shown in FIG. 3,
.uparw..uparw.(t.sub.1,LCA(t.sub.1,t.sub.2))={AND.sub.T,OR}.
[0030] For OR connected patterns, .orgate. denotes that two triples
are related in an OR pattern, i.e., their least common ancestor is
an OR pattern: .orgate.(t,t').ident.LCA(t,t') is OR. In the
example, t.sub.2 and t.sub.3 are .orgate.. For OPTIONAL connected
patterns, {circumflex over (.andgate.)} denotes if one triple is
optional with respect to another, i.e., there is an OPTIONAL
pattern guarding t' with respect to t:
{circumflex over (.andgate.)}(t,t').ident..E-backward.p:p.di-elect
cons..uparw..uparw.(t',t)p is OPTIONAL
In the example, t.sub.6 and t.sub.7 are {circumflex over
(.andgate.)} because t.sub.7 is guarded by an OPTIONAL in relation
to t.sub.6.
[0031] The data flow graph is a graph of G=<V,E>, where
V=(.times.).orgate.root, where root is a special node added to the
graph. A directed edge (t,m).fwdarw.(t',m') exists in V when the
following conditions hold: (t,m) (t',m')(.orgate.(t,t'){circumflex
over (.andgate.)}(t',t)). In addition, a directed edge from root
exists to a node (t,m) if (t,m)=.phi..
[0032] In the example, a directed edge root.fwdarw.(t.sub.4,aco)
exists in the data flow graph (in FIG. 4 the whole graph is shown
and for simplicity in the figure the root node is ommited), because
t.sub.4 can be accessed by an object with a constant. In addition,
it has no required variables. Further,
(t.sub.4,aco).fwdarw.(t.sub.2,aco) is part of the data flow graph,
because (t.sub.2,aco) has a required variable y that is produced by
(t.sub.4,aco). In turn, (t.sub.2,aco) has an edge to (t.sub.1,acs),
because (t.sub.1,acs) has a required variable x which is produced
by (t.sub.2,aco).
[0033] The data flow graph is weighted, and the weights for each
edge between two nodes is determined by a function:
W((t,m),(t',m')),S).fwdarw.w. The w is derived from the costs of
the two nodes, i.e., TMC(t,m,S), and TMC(t',m',S). A simple
implementation of this function, for example, could apply the cost
of the target node to the edge. In the example, for instance, w for
the edge root.fwdarw.(t.sub.4,aco) is 2, whereas the edge
root.fwdarw.(t.sub.4,asc) is .
[0034] Given a weighted data flow graph , the optimal, in terms of
minimizing the cost, order for accessing all the triples in query Q
is the minimal weighted tree that covers all the triples in Q,
which is NP-hard. Since the query can contain a large number of
triples, a greedy algorithm to is used to determine the execution
tree. If T denotes the execution tree that is being computed and
.tau. refers to the set of triples corresponding to nodes already
in the tree,
.tau..ident.{t.sub.i|.E-backward.m.sub.i(t.sub.i,m.sub.i.di-elect
cons.}. The object is to add a node that adds a new triple to the
tree while adding the cheapest possible edge. Formally, a node
(t',m') is chosen such that:
(t',m').di-elect cons.V#node to add
t'.tau.#node adds new triple
.E-backward.(t,m):
( ( t , m ) .di-elect cons. # node to add ( t , m ) .fwdarw. ( t '
, m ' ) # valid edge to new node # no similar node such that ... (
t '' m , m '' ) , ( t ''' , m ''' ) : ( ( t '' , m '' ) .di-elect
cons. t ''' .tau. ( t '' , m '' ) .fwdarw. ( t ''' , m ''' ) #
adding ( t ''' , m ''' ) is cheaper W ( ( t '' , m '' ) , ( t ''' ,
m ''' ) ) < W ( ( t , m ) , ( t ' , m ' ) ) ) ) ##EQU00001##
[0035] On the first iteration, .sub.0=root, and .tau..sub.0=.phi..
.sub.i+1 computed by applying the step defined above, and the
triple of the chosen node is added to .tau..sub.i+1. In the
example, root.fwdarw.(t.sub.4,aco) is the cheapest edge, so
.sub.1=(t.sub.4,aco), and .tau..sub.0=t.sub.4. Then (t.sub.2,aco)
is added to .sub.2, and so on. The iterations stop at .sub.n, where
n is the number of triples in Q. FIG. 4 shows the computed tree, as
indicated by nodes 408, and the algorithm below computes the
optimal slow tree, where function triple(j) returns the triple
associated with a node in .
TABLE-US-00001 Input: The weighted data flow graph G Output: An
optimal flow tree T 1 .tau. .rarw. O; 2 J .rarw. root; 3 E .rarw.
SortEdgesByCost(G); 4 while |J| < |Q| do 5 for each edge
e.sub.ij .di-elect cons. E do 6 if i .di-elect cons. J j J
triple(j) .tau. then 7 J .rarw. J .orgate. j; 8 .tau. .rarw. .tau.
.orgate. triple(j); 9 T .rarw. e.sub.ij;
[0036] Both the data flow graph and the optimal flow tree largely
ignore the query structure, i.e., the organization of triples into
patterns, and the operators between the (triple) patterns. Yet,
they provide useful information as to how to construct an actual
plan for the input query, the focus of this section and output of
the QPB module.
[0037] In more detail, the main algorithm ExecTree of the module
appears below. The algorithm is recursive and takes as input the
optimal flow tree F computed by DFB, and (the parse tree of) a
pattern P, which initially is the main pattern that includes the
whole query.
TABLE-US-00002 Input: The optimal flow tree F of query Q,a pattern
P in Q Output: An execution tree T for P,a set L of execution sub -
trees 1 T .rarw. O; L .rarw. O; 2 switch the type of pattern P do 3
case P is a SIMPLE pattern 4 for each triple pattern t.sub.i
.di-elect cons. P do 5 T.sub.i .rarw. GetTree(t.sub.i,F); L.sub.i
.rarw. O; 6 if isLeaf(T.sub.i,F) then L .rarw. L .orgate. T.sub.i;
7 else (T,L) .rarw. AndTree(F,T,L,T.sub.i,L.sub.i); 8 case P is an
AND pattern 9 for each sub-pattern P.sub.i .di-elect cons. P do 10
(T.sub.i,L.sub.i) .rarw. ExecTree(F,P.sub.i); 11 (T,L) .rarw.
AndTree(F,T,L,T.sub.i,L.sub.i); 12 case P is an OR pattern 13 for
each sub-pattern P.sub.i .di-elect cons. P do 14 (T.sub.i,L.sub.i)
.rarw. ExecTree(F,P.sub.i); 15 .rarw. OrTree(F,T,L,T.sub.i,L.sub.i)
16 case P is an OPTIONAL pattern 17 (T',L') .rarw. ExecTree(F,P);
18 (T,L) .rarw. OptTree(F,T,L,T.sub.i,L.sub.i) 19 case P is a
nested pattern 20 (T,L) .rarw. ExecTree(F,P); 21 return (T,L)
[0038] In the running example, for the query 202 in FIG. 2, the
algorithm takes as input the parse tree in FIG. 3 and the optimal
flow tree in FIG. 4. The algorithm returns a schema-independent
plan T, called the execution tree for the input query pattern P.
The set of returned execution sub-trees is guaranteed to be empty
when the recursion terminates, but contains important information
that the algorithm passes from one level of recursion to the
previous one(s) while the algorithm runs.
[0039] There are four main types of patterns in SPARQL, namely,
SIMPLE, AND, UNION (a.k.a OR), and OPTIONAL patterns, and the
algorithm handles each one independently as illustrated through the
running example. Initially, both the execution tree T and the set
are empty (line 1). Since the top-level node in FIG. 3 is an AND
node, the algorithm considers each sub-pattern of the top-level
node and calls itself recursively (lines 8-10) with each of the
subpatterns as argument. The first sub-pattern recursively
considered is a SIMPLE one consisting of the single triple pattern
t.sub.1. By consulting the flow tree F, the algorithm determines
the optimal execution tree for t.sub.1 which consists of just the
node (t.sub.1,acs) (line 5). By further consulting the flow (line
6) the algorithm determines that node (t.sub.1,acs) is a leaf node
in the optimal flow and therefore it's evaluation depends on the
evaluation of other flow nodes. Therefore, the algorithm adds tree
(t.sub.1,acs) to the local late fusing set of execution trees. Set
contains execution sub-trees that should not be merged yet with the
execution tree T but should be considered later in the process.
Intuitively, late fusing plays two main roles. It uses the flow as
a guide to identify the proper point in time to fuse the execution
tree T with execution sub-trees that are already computed by the
recursion, and it aims to optimize query evaluation by minimizing
the size of intermediate results computed by the execution tree.
Therefore, it only fuses sub-trees at the latest possible place,
when either the corresponding sub-tree variables are needed by the
later stages of the evaluation, or when the operators and structure
of the query enforce the fuse.
[0040] The first recursion terminates by returning
(T.sub.1,.sub.1)=(.phi.,{L.sub.1=(t.sub.1,acs)}). The second
sub-pattern in FIG. 3 is an OR and is therefore handled in lines
12-15. The resulting execution sub-tree contains three nodes, an OR
node as root (from line 15) and nodes (t.sub.2,aco) and
(t.sub.3,aco) as leaves (recursion in line 14). This sub-tree is
also added to local set and the second recursion terminates by
returning
(T.sub.2,.sub.2)=(.phi.,{L.sub.2={OR,(t.sub.2,aco),(t.sub.3,aco)}}).
Finally, the last sub-pattern in FIG. 3 is an AND pattern again,
which causes further recursive calls in lines 8-11. In the
recursive call that processes triple t.sub.4 (lines 5-7), the
execution tree node (t.sub.4,aco) is the root node in the flow and
therefore it is merged to the main execution tree T. Since T is
empty, it becomes the root of the tree T. The three sub-trees that
include nodes (t.sub.5,aco),(t.sub.6,acs), and
OPT={(OPTIONAL),(t.sub.7,aco)} are all becoming part of set .
Therefore, the third recursion terminates by returning
(T.sub.3,.sub.3)=(t.sub.4,aco),{.sub.3={(t.sub.5,aco)},.sub.4={-
(t.sub.6,aco)t.sub.5,aco},.sub.5={(OPTIONAL),(t.sub.7,aco)}}. After
each recursion ends (line 10), the algorithm considers (line 11)
the returned execution T.sub.i and late-fuse .sub.i trees and uses
function AndTree to build a new local execution T and set of
late-fusing trees (by also consulting the flow and following the
late-fusing guidelines on postponing tree fusion unless it is
necessary for the algorithm to progress). So, after the end of the
first recursion and the first call to function AndTree,
(T,)=(T.sub.1,.sub.1), i.e., the trees returned from the first
recursion. After the end of the second recursion, and the second
call to AndTree, (T,)=(.phi.,.sub.1.orgate..sub.2). Finally, after
the end of the third recursion,
(T,)=((t.sub.4,aco),.sub.1.orgate..sub.2.orgate..sub.3).
[0041] The last call to AndTree builds the tree 500 illustrated in
of FIG. 5 as follows. Starting from node (t.sub.4,aco), it consults
the flow and picks from the set the sub-tree L.sub.2 and connects
this to node (t.sub.4,aco) by adding a new AND node as the root of
the tree. Sub-trees L.sub.3, L.sub.4 and L.sub.5 can be added at
this stage to T but they are not considered as they violate the
principles of late-fusing (their respective variables are not used
by any other triple, as is also obvious by the optimal flow). On
the other hand, there is still a dependency between the latest tree
T and L.sub.1 since the selectivity of t.sub.1 can be used to
reduce the intermediate size of the query results (especially the
bindings to variable ? y). Therefore, a new AND is introduced and
the existing T is extended with L.sub.1. The process iterates in
this fashion until the whole tree in FIG. 5 is generated.
[0042] Using the optimal flow tree as a guide enabled weavinge the
evaluation of different patterns, while the structured based
processing guarantees that the associativity of operations in the
query is respected. So, the optimizer can generate plans like the
one in FIG. 5 where only a portion of a pattern is initially
evaluated, e.g., node (t.sub.4,aco), while the evaluation of other
constructs in the pattern, e.g., node (t.sub.5,aco), can be
postponed until it no longer can be avoided. At the same time, this
de-coupling from query structure facilitates pushing the evaluation
of patterns early in the plan, e.g., node (t.sub.1,acs), when doing
so improves selectivity and reduces the size of intermediate
results.
[0043] The SPARQL to SQL translator takes as input the execution
tree generated from the QPB module and performs two operations.
First, it transforms the execution tree into an equivalent query
plan that exploits the entity-oriented storage of, for example,
R2DF. Second, it uses the query plan to create the SQL query which
is executed by the database.
[0044] In order to build the query plan, the execution tree
provides an access method and an execution order for each triple
but assumes that each triple node is evaluated independently of the
other nodes. However, one of the advantages of the entity-oriented
storage is that a single access to, say, the DPH relation might
retrieve a row that can be used to evaluate multiple triple
patterns (star-queries). To this end, starting from the execution
tree the translator builds a query plan where triples with the same
subject (or the same object) are merged in the same plan node. A
merged plan node indicates to the SQL builder that the containing
triples form a star-query and is executed with a single SQL select.
Merging of nodes is always advantageous with one exception, when
the star query involves entities with spills. The presence of such
entities would require self-joins of the DPH (RPH) relations in the
resulting SQL statement. Self-joins are expensive, and therefore
the following strategy is used to avoid them. When the star-queries
involve entities with spills, the evaluation of the star-query is
cascaded by issuing multiple SQL statements, each evaluating a
subset of the star-query while at the same time filtering entities
from the subsets of the star-query that have been previously
evaluated. The multiple SQL statements are such that no SQL
statement accesses predicates stored into different spill rows. Of
course, the question remains on how to determine whether spills
affect a star query. In accordance with exemplary embodiment of the
methods and systems of the present invention, this is
straightforward. With only a tiny fraction of predicates involved
in spills, e.g., due to coloring, the optimizer consults an
in-memory structure of predicates involved in spills to determine
during merging whether any of the star-query predicates participate
in spills.
[0045] During the merging process, both the structural and semantic
constraints are respected. The structural constraints are imposed
by the entity-oriented representation of data. To satisfy the
structural constraints, candidate nodes for merging need to refer
to the same entity, have the same access method and do not involve
spills. As an example, in FIG. 4 nodes t.sub.2 and t.sub.3 refer to
the same entity due to variable ? x and the same access method aco,
as do nodes t.sub.6 and t.sub.7, due to the variable ? y and the
method acs.
[0046] Semantic constraints for merging are imposed by the control
structure of the SPARQL query, i.e., the AND, UNION, OPTIONAL
patterns. This restricts the merging of triples to constructs for
which we can provide the equivalent SQL statements to access the
relational tables. Triples in conjunctive and disjunctive patterns
can be safely merged because the equivalent SQL semantics are well
understood. Therefore, with a single access, the system, can check
whether the row includes the non-optional predicates in the
conjunction. Similarly, it is possible to check the existence of
any of the predicates mentioned in the disjunction. More formally,
to satisfy the semantic constraints of SPARQL, candidate nodes for
merging need to be ANDMergeable, ORMergeable or OPTMergeable.
[0047] For AND Merge able nodes, two nodes are ANDMergeable if f
their least common ancestor and all intermediate ancestors are AND
nodes: ANDMergeable(t,t').A-inverted.x:x.di-elect
cons.(.uparw..uparw.(t,LCA(t,t')).orgate..uparw..uparw.(t',LCA(t,t')))x
is AND. For OR Merge able nodes, two nodes are ORMergeable if f
their least common ancestor and all intermediate ancestors are OR
nodes: ORMergeable(t,t').A-inverted.x:x.di-elect
cons.(.uparw..uparw.(t,LCA(t,t')).orgate..uparw..uparw.(t',LCA(t,t')))x
is OR. Going back to the execution tree in FIG. 5, notice that
ORMergeable (t.sub.2,t.sub.3) is true, but ORMergeable
(t.sub.2,t.sub.5) is false. For OPTIONAL Merge able nodes, two
nodes are OPTMergeable if f their least common ancestor and all
intermediate ancestors are AND nodes, except the parent of the
higher order triple in the execution plan which is OPTIONAL:
OPTMergeable(t,t').A-inverted.x:x.di-elect
cons.(.uparw..uparw.(t,LCA(t,t')).orgate..uparw..uparw.(t',LCA(t,t')))x
is AND{x is OPTIONAL x is parent of t'}. As an example, in FIG. 5
OPTMergeable (t.sub.6,t.sub.7) is true.
[0048] Given the input execution tree, pairs of nodes are
identified that satisfy both the structural and semantic
constraints introduced and are merged. So, given as input the
execution tree in FIG. 5, the resulting query plan tree 600 is
shown in FIG. 6. Notice that in the resulting query plan there are
two node merges, one due to the application of the ORMergeable
definition, and one by the application of the OPTMergeable
definition. Each merged node is annotated with the corresponding
semantics under which the merge was applied. As a counter-example,
consider node (t.sub.5,aco) which is compatible structurally with
the new node ({t.sub.2,t.sub.3}, aco) since they both refer to the
same entity through variable ? y, and have the same access method
aco. However, these two nodes are not merged since they violate the
semantic constraints, i.e., they do not satisfy the definitions
above since their merge would mix a conjunctive with a disjunctive
pattern. Even for the simple running example, the two identified
node merges result in significant savings in terms of query
evaluation. Intuitively, one can think of these two merges as
eliminating two extra join operations during the translation of the
query plan to an actual SQL query over the database schema.
[0049] SQL generation is the final step of query translation. The
query plan tree plays an important role in this process, and each
node in the query plan tree, be it a triple, merge or control node,
contains the necessary information to guide the SQL generation. For
the generation, the SQL builder performs a post order traversal of
the query plan tree and produces the equivalent SQL query for each
node. The whole process is assisted by the use of SQL code
templates.
[0050] In more detail, the base case of SQL translation considers a
node that corresponds to a single triple or a merge. FIG. 7 shows
the template used to generate SQL code for such a node. The code in
box 1 702 sets the target of the query to the DPH or RPH tables,
according to the access method in the triple node. The code in box
2 704 restricts the entities being queried. As an example, when the
subject is a constant and the access method is acs, the entry is
connected to the constant subject values. When the subject is
variable and the method is acs, then entry is connected with a
previously-bound variable from a prior SELECT sub-query. The same
reasoning applies for the entry component for an object when the
access method is aco. Box 3 706 illustrates how one or more
predicates are selected. That is, when the plan node corresponds to
a merge, multiple pred.sub.i components are connected through
conjunctive or disjunctive SQL operators. Finally, box 4 708 shows
how to do outer join with the secondary table for multi-valued
predicates.
[0051] The operator nodes in the query plan are used to guide the
connection of instantiated templates like the one in FIG. 7. It has
already been shown how AND nodes are implemented through the
variable binding across triples as in box 2. For OR nodes we use
the SQL UNION operator to connect its components' previously
defined SELECT statements. For OPTIONAL, LEFT OUTER JOIN is used
between the SQL template for the main pattern and the SQL template
for the OPTIONAL pattern. The final SQL for the running example
where the SQL templates described above are instantiated according
to the query plan tree in FIG. 6 and the SPARQL query of FIG. 2
appears below. [0052] WITH QT.sub.4RPH AS [0053] SELECT
T.val.sub.1ASval.sub.1FROM RPH AS T WHERE T.entry=`Software`AND
T.pred.sub.1=`industry`. [0054] QT.sub.4DS AS [0055] SELECT
COALESCE (S.elm,T.val.sub.1) AS y [0056] FROM QT.sub.44RPH AS T
LEFT OUTER JOIN DS AS S ON T.val.sub.1=S.l_id [0057] QT.sub.23RPH
AS [0058] SELECT QT.sub.4DS.y, [0059] CASE T.pred.sub.m=`founder`
THEN val.sub.m ELSE null END AS val.sub.m, [0060] CASE
T.pred.sub.0=`member` THEN val.sub.0 ELSE null END AS val.sub.0
[0061] FROM RPH AS T, QT.sub.4DS [0062] WHERE T.entry=QT.sub.4DS.y
AND (T.pred.sub.m=`founder` OR T.pred.sub.0=`member`), [0063]
QT.sub.23 AS [0064] SELECT LT.val.sub.0 AS x,T.y FROM QT.sub.23RPH
as T,TABLE (T.val.sub.m,T.val.sub.0) as LT(val.sub.0) [0065] WHERE
LT.val.sub.0 IS NOT NULL [0066] QT.sub.1DPH AS [0067] SELECT
T.entry AS x,QT.sub.23.y FROM DPH AS T,QT.sub.23 [0068] WHERE
T.entry=QT.sub.23.x AND T.pred.sub.k=`home` AND T.val.sub.1=`Palo
Alto`, [0069] QT.sub.5RPH AS [0070] SELECT T.entry AS
y,QT.sub.1DPH.x FROM RPH AS T,QT.sub.1DPH [0071] WHERE
T.entry=QT.sub.1DPH.y AND T.pred.sub.1=`developer`, [0072]
QT.sub.67DPH AS [0073] SELECT T.entry AS y,QT.sub.5RPH.x,CASE
T.pred.sub.k=`employees` THEN val.sub.k ELSE null END as z [0074]
FROM DPH AS T,QT.sub.5RPH WHERE T.entry=QT.sub.5RPH.y AND
T.pred.sub.m=`revenue` [0075] SELECT x,y,z FROM QT.sub.67DPH
[0076] As illustrated above, several Common Table Expressions
(CTEs) are used for each plan node. For example, t.sub.4 is
evaluated first and accesses RPH using the Software constant. Since
industry is a multivalued predicate, the RS table is also accessed.
The remaining predicates in this example are single valued and the
access to the secondary table is avoided. The ORMergeable node
t.sub.23 is evaluated next using the RPH table where the object is
bound to the values of y produced by the first triple. The WHERE
clause enforces the semantic that at least one of the predicates is
present. The CTE projects the values corresponding to the present
predicates and null values for those that are missing. The next CTE
just flips these values, creating a new result record for each
present predicate. The plan continues with triple t.sub.5 and is
completed with node the OPTMergeable node t.sub.67. Here no
constraint is imposed for the optional predicate making its
presence optional on the record. In case the predicate is present,
the corresponding value is projected, otherwise null. In this
example, each predicate is assigned to a single column. When
predicates are assigned to multiple columns, the position of the
value is determined with CASE statements as seen in the SQL
sample.
[0077] To examine the effectiveness of the query optimization,
experiments were conducted using both a 1M triple microbenchmark
and queries from other datasets. As an example, for the
microbenchmark, two constant values O.sub.1 and O.sub.2 were
considered with relative frequency in the data of 0.75 and 0.01,
respectively. Then, the simple query 802 shown in FIG. 8 was used
that allowed data flows in either direction, i.e., evaluation could
start on t.sub.1 with an aco using O.sub.1, then the bindings for ?
s to were used access t.sub.2 with an acs, or start instead on
t.sub.2 with an aco using O.sub.2 and use bindings for ? s to
access t.sub.1. The latter case is better.
[0078] In FIG. 8, the SQL generated by our SPARQL optimizer 804 is
shown, while an equivalent SQL query corresponding to the only
alternative but sub-optimal flow 806 is also shown. The former
query took 13 ms to evaluate, whereas the latter took 5 times
longer, i.e., 65 ms, suggesting that the optimization is in fact
effective even in this simple query. Using real and benchmark
queries from datasets resulted in even more striking differences in
evaluation times. For example, when optimized by our SPARQL
optimizer query, PQ1 from PRBench (Section 4) was evaluated in 4
ms, while the translated SQL corresponding to a sub-optimal flow
required 22.66 seconds.
[0079] The performance of an R2DF schema, using IBM DB2 as the
relational back-end, was compared to that of Virtuoso 6.1.5
OpenSource Edition, Apache Jena 2.7.3 (TDB), OpenRDF Sesame 2.6.8,
and RDF-3X 0.3.5. R2DF, Virtuoso and RDF-3X were run in a client
server mode on the same machine and all other systems were run in
process mode. For both Jena and Virtuoso, all recommended
optimizations were enabled. Jena had the BGP optimizer enabled. For
Virtuoso all recommended indexes were built. For R2DF, only indexes
on the entry columns of the DPH and RPH relations were added, (no
indexes on the pred.sub.i and val.sub.i columns).
[0080] Experiments were conducted with 4 different benchmarks,
LUBM, SP2Bench, DBpedia, and a private benchmark PRBench. The LUBM
and SP2Bench benchmarks were scaled up to 100 million triples each,
and their associated published query workloads were used. The
DBpedia 3.7 benchmark has 333 million triples. The private
benchmark included data from a tool integration application, and it
contained 60 million triples about various software artifacts
generated by different tools, e.g., bug reports, requirements, etc.
For all systems, queries were evaluated in a warm cache scenario.
For each dataset, benchmark queries were randomly mixed to create a
run, and each run was issued 8 times to the 5 stores. The first run
was discard, and the average result for each query over 7
consecutive runs was reported. For each query, its running time was
measured excluding the time taken to stream back the results to the
API, in order to minimize variations caused by the various APIs
available. As shown in Table 1, the evaluated queries were
classified into four categories. Queries that failed to parse
SPARQL correctly were reported as unsupported. The remainder
supported queries were further classified as either complete,
timeout, or error. The results from each system were counted, and
when a system provided the correct number of answers, the query was
classified as completed. If the system returned the wrong number of
results, this was classified as an error. Finally, a timeout of 10
minutes was used to trap queries that do not terminate within a
reasonable amount of time. In the table, the average time taken (in
seconds) to evaluate complete and timeout queries is reported. For
queries that timeout, their running time was set to 10 minutes. The
time of queries that return the wrong number of results is not
reported.
[0081] This is the most comprehensive evaluation of RDF systems.
Unlike previous works, this is the first study that evaluates 5
systems using a total of 78 queries, over a total of 600 million
triples. The experiments were conducted on 5 identical virtual
machines (one per system), each equivalent to a 4-core, 2.6 GHz
Intel Xeon system with 32 GB of memory running 64-bit Linux. Each
system was not memory limited, meaning it could consume all of its
32 G. None of the systems came close to this memory limit in any
experiment.
TABLE-US-00003 TABLE 1 Summary Results for All Systems and Datasets
Supported Unsu- Mean Dataset System Complete Timeout Error pported
(secs) LUBM Jena 12 -- -- -- 35.1 (100M Sesame 4 -- 8 -- 164.7
triples) Virtuoso 12 -- -- -- 16.8 (12 queries) RDF-3X 11 -- -- 1
2.8 R2DF 12 -- -- -- 8.3 SP2Bench Jena 11 6 -- -- 253 (100M Sesame
8 8 1 -- 330 triples) Virtuoso 16 1 -- -- 211 (17 queries) RDF-3X 6
2 2 7 152 R2DF 16 1 -- -- 108 DBpedia Jena 18 1 1 -- 33 (333M
Virtuoso 20 -- -- -- 0.25 triples) R2DF 20 -- -- -- 0.25 (20
queries) PRBench Jena 29 -- -- -- 5.7 (60M triples) Virtuoso 25 --
-- 4 3.9 (29 queries) R2DF 29 -- -- -- 1.0
[0082] The LUBM benchmark requires OWL DL inference, which is not
supported across all tested systems. Without inference, most
benchmark queries return empty result sets. To address this issue,
the existing queries were expanded, and a set of equivalent queries
that implement inference and do not require this feature from the
evaluated system was created. As an example, if the LUBM ontology
stated that GraduateStudent Student, and the query asks for ? x
rdf:type Student, the query was expanded into ? x rdf:type Student
UNION ? x rdf:type Graduate Student. This set of expansions was
performed, and the same expanded query was issued to all systems.
From the 14 original queries in the benchmark, only 12 (denoted as
LQ1 to LQ1Q,LQ13 and LQ14) are included here because 2 queries
involved ontological axioms that cannot be expanded.
[0083] SP2Bench is an extract of DBLP data with corresponding
SPARQL queries (denoted as SQ1 to SQ17). This benchmark was used as
is, with no modifications. Prior reports on this benchmark were
conducted with at most 5 million triples. It was scaled to 100
million triples, and some queries (by design) had rather large
result sets. SQ4 in particular created a cross product of the
entire dataset, which meant that all systems timeout on this
query.
[0084] The DBpedia SPARQL benchmark is a set of query templates
derived from actual query logs against the public DBpedia SPARQL
endpoint. These were used templates with the DBpedia 3.7 dataset,
and 20 queries (denoted as DQ1 to DQ20) were obtained that had
non-empty result sets. Since templates were derived for an earlier
DBpedia version, not all result in non-empty queries.
[0085] The private benchmark reflects data from a tool integration
scenario where specific information about the same software
artifacts are generated by different tools, and RDF data provides
an integrated view on these artifacts across tools. This is a quad
dataset where triples are organized into over 1 million `graphs`.
This caused problems for some systems which do not support quads,
e.g., RDF-3X, Sesame. Twenty nine SPARQL queries (denoted as PQ1 to
PQ29) were used, with some being fairly complex queries, e.g., a
SPARQL union of 100 conjunctive queries.
[0086] Table 1 shows that R2DF is the only system that evaluates
correctly and efficiently 77 out of the 78 tested queries. As
mentioned, SQ4 was the only query in which the system did timeout
as did all the other systems. If SQ4 is excluded, it is clear from
the table that each of the remaining systems had queries returning
incorrect number of results, or queries that timeout without
returning any results. The advantage of R2DF is not emphasized in
terms of SPARQL support, since this is mostly a function of system
maturity and continued development.
[0087] Given Table 1, it is hard to make direct system comparisons.
Still, when the R2DF system is compared with systems that can
evaluate approximately the same queries, i.e., Virtuoso and Jena,
then R2DF is in the worst case slightly faster, and in the best
case, as much as an order of magnitude faster than the other two
systems. So, for LUBM, R2DF is significantly faster than Virtuoso
(2.times.) and Jena (4.times.). For SP2Bench, R2DF is on average
times about 50% faster than Virtuoso, although Virtuoso has a
better geometric mean (not shown due to space constraints), which
reflects Virtuoso being much better on short running queries. For
DBpedia, R2DF and Virtuoso have comparable performance, and for
PRBench, R2DF is about 5.5.times. better than Jena. Jena is
actually the only system that supports the same queries as R2DF,
and across all datasets R2DF is in the worst case 60%, and in the
best case as much as two orders of magnitude faster. A comparison
between R2DF and RDF-3X is also possible, but only in the LUBM
dataset where both systems support a similar number of queries. The
two systems are fairly close in performance and out-perform the
remaining three systems. When compared between themselves across 11
queries (RDF-3X did not run one query), R2DF is faster than RDF-3X
in 3 queries, namely in LQ8, LQ13 and LQ14 (246 ms, 14 ms and 4.6
secs versus 573 ms, 36 ms and 9.5 secs, respectively), while RDF-3X
has clearly an advantage in 3 other queries, namely in LQ2, LQ6,
LQ10 (722 ms, 12 secs and 1.57 secs versus 20 secs, 33 secs and
3.42 secs, respectively). For the remaining 5 queries, the two
systems have almost identical performance with RDF-3X being faster
than R2DF by approximately 3 ms for each query.
[0088] For a more detailed per-query comparison, FIG. 9 illustrates
the running times for R2DF, Virtuoso and Jena for all 12 LUBM
queries. Reported times are in milliseconds and the scale is
logarithmic). Notice that R2DF outperforms the other systems in the
long-running and complicated queries (e.g., LQ6, LQ8, LQ9, LQ13,
LQ14). So, R2DF takes approximately 34 secs to evaluate LQ6, while
Virtuoso requires 83.2 secs and Jena 150 secs. Similarly, R2DF
takes 40 secs to evaluate LQ9, whereas Virtuoso requires 46 and
Jena 60 secs. Most notably, in LQ14 R2DF requires 4.6 secs while
Virtuoso requires 53 secs and Jena 94.1 sees. For the sub-second
queries, R2DF is slightly slower than the other systems, but the
difference is negligible at this scale. So, for LQ1, R2DF requires
5 ms, while Virtuoso requires 1.8 ms and Jena 2.1 ms. Similarly,
for LQ3 R2DF requires 3.4 ms while Virtuoso takes 1.8 ms and Jena
2.0 ms.
[0089] The situation is similar in the PRBench case. FIG. 10 shows
the evaluation time of 4 long-running queries. Consistently, R2DF
outperforms all other systems. For example, for PQ10 R2DF takes 3
ms, while Jena requires 27 seconds and Virtuoso requires 39
seconds. For each of the other three queries, R2DF takes
approximately 4.8 secs while Jena requires a minimum of 32 and
Virtuoso a minimum of 11 secs. FIG. 11 shows that the situation is
similar for medium-running queries where R2DF consistently
outperforms the competition.
[0090] Methods and systems in accordance with exemplary embodiments
of the present invention can take the form of an entirely hardware
embodiment, an entirely software embodiment or an embodiment
containing both hardware and software elements. In a preferred
embodiment, the invention is implemented in software, which
includes but is not limited to firmware, resident software and
microcode. In addition, exemplary methods and systems can take the
form of a computer program product accessible from a
computer-usable or computer-readable medium providing program code
for use by or in connection with a computer, logical processing
unit or any instruction execution system. For the purposes of this
description, a computer-usable or computer-readable medium can be
any apparatus that can contain, store, communicate, propagate, or
transport the program for use by or in connection with the
instruction execution system, apparatus, or device. Suitable
computer-usable or computer readable mediums include, but are not
limited to, electronic, magnetic, optical, electromagnetic,
infrared, or semiconductor systems (or apparatuses or devices) or
propagation mediums. Examples of a computer-readable medium include
a semiconductor or solid state memory, magnetic tape, a removable
computer diskette, a random access memory (RAM), a read-only memory
(ROM), a rigid magnetic disk and an optical disk. Current examples
of optical disks include compact disk--read only memory (CD-ROM),
compact disk--read/write (CD-R/W) and DVD.
[0091] Suitable data processing systems for storing and/or
executing program code include, but are not limited to, at least
one processor coupled directly or indirectly to memory elements
through a system bus. The memory elements include local memory
employed during actual execution of the program code, bulk storage,
and cache memories, which provide temporary storage of at least
some program code in order to reduce the number of times code must
be retrieved from bulk storage during execution. Input/output or
I/O devices, including but not limited to keyboards, displays and
pointing devices, can be coupled to the system either directly or
through intervening I/O controllers. Exemplary embodiments of the
methods and systems in accordance with the present invention also
include network adapters coupled to the system to enable the data
processing system to become coupled to other data processing
systems or remote printers or storage devices through intervening
private or public networks. Suitable currently available types of
network adapters include, but are not limited to, modems, cable
modems, DSL modems, Ethernet cards and combinations thereof.
[0092] In one embodiment, the present invention is directed to a
machine-readable or computer-readable medium containing a
machine-executable or computer-executable code that when read by a
machine or computer causes the machine or computer to perform a
method for optimizing semantic web queries in accordance with
exemplary embodiments of the present invention and to the
computer-executable code itself. The machine-readable or
computer-readable code can be any type of code or language capable
of being read and executed by the machine or computer and can be
expressed in any suitable language or syntax known and available in
the art including machine languages, assembler languages, higher
level languages, object oriented languages and scripting languages.
The computer-executable code can be stored on any suitable storage
medium or database, including databases disposed within, in
communication with and accessible by computer networks utilized by
systems in accordance with the present invention and can be
executed on any suitable hardware platform as are known and
available in the art including the control systems used to control
the presentations of the present invention.
[0093] While it is apparent that the illustrative embodiments of
the invention disclosed herein fulfill the objectives of the
present invention, it is appreciated that numerous modifications
and other embodiments may be devised by those skilled in the art.
Additionally, feature(s) and/or element(s) from any embodiment may
be used singly or in combination with other embodiment(s) and steps
or elements from methods in accordance with the present invention
can be executed or performed in any suitable order. Therefore, it
will be understood that the appended claims are intended to cover
all such modifications and embodiments, which would come within the
spirit and scope of the present invention.
* * * * *